[VMonG5K] Issues in Monitoring stack deployment - Grafana Timeout
Hi,
I have some issues with deployment on VM on top of G5K (using VMonG5K
provider).
Deployment phase "Waiting for grafana to be ready" stays in timeout. When I log (through ssh) into the VM and run docker ps
, I don't see InfluxDB or Grafana container (same thing with the -a
flag).
Deployment seems OK when I tried a super minimal deployment with only 1 machine (telegraf agent deployed into the same machine as Grafana and InfluxDB).
Here a (somehow) minimal example:
from enoslib.infra.enos_vmong5k.configuration import Configuration as Vmong5k_Configuration
from enoslib.infra.enos_vmong5k.provider import VMonG5k
from enoslib.api import discover_networks
from enoslib.service import Monitoring
import yaml
raw_config = """
deployment:
job_name: BCTMark-vmong5k
image: "/grid5000/virt-images/debian10-x64-base.qcow2"
walltime: "07:50:00"
resources:
machines:
- roles: [dashboard]
cluster: paravance
number: 1
flavour: medium
extra_devices: ""
- roles: [agent]
cluster: paravance
number: 1
flavour: medium
extra_devices: ""
networks: ["ntw_monitoring"]
"""
config = yaml.safe_load(raw_config)["deployment"]
provider = VMonG5k(Vmong5k_Configuration.from_dictionnary(config))
roles, networks = provider.init(force_deploy=True)
roles = discover_networks(roles, networks)
m = Monitoring(collector=roles["dashboard"],
ui=roles["dashboard"],
agent=roles["agent"],
network='ntw_monitoring')
m.deploy()
I think I may be doing something wrong but not sure what.
Expected behavior:
- InfluxDB and Grafana containers up and their ports used
Behavior that I have:
- No error at the InfluxDB deployment but can't find the container
- Verification of Grafana deployment in timeout
Do you have the same issues that I do with the provided code?